DOCUMENTO DE INVESTIGACIÓN
N.º 002 | 2016
Gains from trade:
The Costa Rican case
Alonso Alfaro Ureña
Abril, 2016
1
Las ideas expresadas en estos documentos son de los autores y no necesariamente
representan las del Banco Central de Costa Rica.
La serie Documentos de Investigación del Banco Central de Costa Rica en versión PDF se
encuentra en http://www.bccr.fi.cr
Fotografía de portada: “Los presentes”, conjunto escultórico de nueve figuras en bronce,
año 1983, del artista costarricense Fernando Calvo Sánchez. Colección del Banco Central
de Costa Rica.
Gains from trade:
The Costa Rican case
Alonso Alfaro Ureña
‡
Abstract
One of the oldest and most interesting questions in the economic literature is how to quantify the gains
from trade. Recently, Costinot & Rodrı́guez-Clare (2013) (CRC) developed a methodology that uses
the World Input Output Database (WIOD) to compute this value for a list of countries. Costa Rica has
never been part of this database given the lack of appropriate data. However, with the publication of
a new Input Output Table for Costa Rica, the Foreign Trade Ministry (COMEX) was able to develop
a domestic version of the WIOD that includes the country. This allows for counterfactual exercises in
which we compare the current situation with autarky and other average tariff levels. As CRC show, this
can be done using different productive structures and competition schemes in the economy. The results
can provide valuable information on how much a small open economy like Costa Rica’s can benefit from
international trade, and what are the differences in the results when compared to similar countries.
Key Words: Gains from trade, Costa Rica, input-output.
JEL Classification: F10, I30, D57
Abstract
Uno de los retos más interesantes de la literatura económica es la cuantificación de las ganancias del
Comercio. Recientemente, Costinot & Rodrı́guez-Clare (2013) (CRC) desarrollaron una metodologı́a
que utiliza la World Input Output Database (WIOD) para calcular este valor para una lista de paı́ses.
Costa Rica nunca ha sido parte de esta base de datos debido a que no contaba con los datos apropiados.
Sin embargo, la publicación de la nueva Matriz Insumo Producto para Costa Rica permitió al Ministerio
de Comercio Exterior (COMEX) desarrollar una versión de la WIOD que incluye al paı́s. Esto permite
realizar ejercicios contrafactuales en los que se compara la situación vigente con autarquı́a y otros niveles
de tarifas promedio. Como CRC muestran, esto se puede hacer considerando diferentes estructuras productivas y esquemas de competencia. Los resultados pueden constituir información valiosa sobre cuánto
puede ganar una economı́a pequeña y abierta como la costarricense gracias al comercio internacional, y
en qué se diferencia de otros paı́ses en una situación similar.
Palabras clave: Ganancias del comercio, Costa Rica, insumo-producto.
Clasificación JEL: F10, I30, D57
‡
Departamento de Investigación Económica, División Económica, BCCR. Email : [email protected]
Gains from trade:
The Costa Rican case
Contents
1 Introduction
1
2 The new database
2
3 Computing gains from trade
3
3.1
Armington Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
3.2
Generalization of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
3.3
One sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.4
Multiple sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.5
Tradable intermediate goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8
4 Counterfactual exercises
9
4.1
Gains from trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
4.2
Effects of increases in tariffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5 Conclusions
15
Appendices
18
A Description of the 16 sectors and trade elasticity used in estimation.
18
List of Figures
1
Gains from a unilateral increase in the tariff . . . . . . . . . . . . . . . . . . . . . . . . 13
List of Tables
1
Gains from trade as percentages of income . . . . . . . . . . . . . . . . . . . . . . . . . 12
2
Losses from a simultaneous 40% tariff increase. Multiple sectors, intermediate goods,
heterogeneous firms (Melitz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3
Description of the 16 sectors used in estimation, and trade elasticity used. . . . . . . . . 18
ii
Gains from trade:
The Costa Rican case1
1
Introduction
One of the oldest and most interesting questions in the economic literature is how to quantify the gains
from trade.2 Recent work by Costinot & Rodrı́guez-Clare (2013) (CRC) described how the results of
a wide array of trade models developed in the last two decades can provide parsimonious measures
of the gains from trade. Those include, for example, one sector models, multiple sector models, and
models with intermediate goods. Different structures for how competition works in those markets are
also considered, such as perfect, Bertrand, and monopolistic competition.3
The results presented in CRC4 are useful for evaluating the effects of globalization and the differences
that arise for different countries depending on the level of integration to the rest of the world. The
authors use the World Input Output Database (WIOD) constructed by Dietzenbacher, Los, Stehrer,
Timmer & de Vries (2013) for computing the gains from trade. However, this database does not include
Costa Rica as an individual country, it is included as part of the “Rest of the World”.
For a small, open economy such as Costa Rica it is of particular interest to quantify how much does
the country gain from having its economy open to trade with the rest of the world. Recent work by
Bullón, Mena, Meng, Sanchez, Vargas & Inomata (2015) allowed them to quantify the integration of
the country to Gloval Value Chains (GVCs). Their work allowed for the publication of a domestic version
of the WIOD that includes Costa Rica as a single country in this database, and not part of the Rest of
1
I want to thank Andrés Rodrı́guez-Clare for the discussions and clarifications.
The positive implications of opening to trade have been well established theoretically for decades, see Samuelson
(1939).
3
In all of these models simple expressions can be summarize how much does real consumption increase when the country
opens to trade.
4
This paper focuses exclusively on the computation of the gains from trade from the gravity models presented at the
beginning of CRC. There are further theoretical and empirical discussions in this paper which are not discussed here, but
should be of interest of any reader who wants to understand other modeling options, and the discussion of the costs of
using parsimonious models such as the ones discussed in this paper. These costs are usually related to the restricting
assumptions regarding functional forms that must be used, which may not be a good fit in all dimensions of the data.
2
1
the World. The goal of this paper is to present the results of the CRC methodology using this version
of the WIOD to compute gains from trade for the Costa Rican economy.
The results are, in general, consistent with the gains from trade from similar small open economies.
The gains from the current situation are above the average of the rest of the world, while incresing
dramatically when the assumptions allow for multiple sectors in perfect competition.
2
The new database
Costa Rica did not update its own Input Output Matrix (IOM) for many decades. Leiva & Vargas
(2014) mention that before 2014 there had been only two matrices in the history of the country, one
from 1969 (Modelo Insumo-Producto para Costa Rica - 1969: Un ensayo de Economı́a Inter-industrial),
and the 2011 version developed by the Banco Central de Costa Rica (BCCR). There have been other
approximations in between, such as the matrix from 1991, which had been the most widely used before
the new publication. Even though in February 2016 the newest version of the IOM (with data for 2012)
was published by BCCR along with a new set of data for the national accounts, this matrix has not
been included in a newer version of the WIOD.
The 2011 version which was published in 2014 was constructed using the most recently available information in accordance with the best practices recommended by the United Nations Statistical Commision.
In Bullón et al. (2015) the authors document how this IOM was embedded into the World Input Output Database. The authors of Bullón et al. (2015), the Ministry of Foreign Trade (COMEX) and the
Central Bank (BCCR) deserve recognition for preparing this database for external use. There is a significant amount of work that can be done thanks to this effort, and the trade and industrial organization
literature of Costa Rica can expand much more thanks to this accomplishment.
The Costa Rican 2011 IOM has 76 products, which were aggregated into 35 industries to match the
international version. The results shown in this paper are not exactly the same as those presented in
Costinot & Rodrı́guez-Clare (2013) because the version into which the Costa Rican IOM was embedded
was the 2011, whereas the authors use the 2008 version. It is also the case that this database shows
trade data after the 2008-2009 crisis that caused a collapse of the quantity of international trade in the
following years, which affects the magnitude of the gains from trade.
One relevant difference from CRC is that for the calculations presented in this paper, 16 sectors are
used for the aggregation levels, instead of the 31 sectors used in the original paper. The reason for this
is that the Costa Rican IOM lacks data on some of the sectors, and makes the inversion of the matrices
required for the computation impossible without some additional aggregation.
2
3
Computing gains from trade
As it was mentioned, the goal of this paper is to apply the same methodology of CRC for the version of
the WIOD that includes Costa Rica as a separate country (CRWIOD henceforth), which was developed
by COMEX. Even though there is no new theory developed throughout this paper, some basic elements
of how the gains from trade are computed in each version will be described explicitly.
I present the main elements that help to understand how the model described works: the preferences, the
price index that corresponds to those preferences, the price of each good according to the assumption
made regarding the competition of the economy, and the gravity equation that results from the solution
of the model. While I give general notions of the relevant elements of each model, the complete
description of the derivation can be found in CRC. The discussion of the many caveats that should be
considered when analyzing each type of model are also found in that paper.
3.1
Armington Model
The simplest multi-country gravity model used in international trade which can match trade patterns
across countries is an Armington model, which assumes an endowment economy. This setup can serve
as a benchmark for comparison for the rest of the models and assumptions presented in the rest of the
paper. In each of the i = 1, . . . , n countries there is an endowment of a unique domestic good Qi . The
preferences take the form
Cj =
n
X
1−σ
σ
σ−1
σ
!
σ
σ−1
ψij Cij
,
(3.1)
i=1
with Cij is the demand for a good exported by country denoted with the first subscript, in this case
i, to the country denoted with the second subscript, j for this example. The parameters ψij > 0 are
exogenous preference parameters, and σ > 1 is the elasticity of substitution between the goods. There
is a price index associated with the goods consumed in each country j:
Pj =
n
X
!1−σ
1−σ 1−σ
ψij
Pij
.
(3.2)
i=1
where Pij is the price of the good produced in country i (also called “good i” because it is endowed to
that country) exported to country j. The trade costs τij of a good being exported from country i to
3
country j are assumed to take an iceberg form:
Pij =
Yi τij
Qi
(3.3)
where Yi is country’s i income, and Qi represents the endowment of the good i. This simple economic
environment results in a gravity equation for the trade flows between each pair of countries, Xij , which
takes the form:5
(Yi τij )−ε χij
Ej
Xij = Pn
−ε
l=1 (Yl τlj ) χlj
(3.4)
where ε, is the trade elasticity,
∂ ln(Xij /Xjj )/∂ ln τij = ε = σ − 1.
Arkolakis, Costinot & Rodriguez-Clare (2012) show that for a wide variety of trade models it is possible to
compute the changes in real consumption when comparing steady state equilibriums6 from two sufficient
statistics, namely the elasticity of imports with respect to the variable trade costs, ε, and the share of
expenditure in domestic goods, λjj =
Xjj
Ej .
Here λjj is formally defined as the share of expenditure on
goods from the same country.
n
λjj =
X
X
Xjj
=1−
Xij /
Xij .
Ej
(3.5)
i=1
i6=j
For these types of models, welfare changes are defined as changes in real consumption given a foreign
shock. In the case of the Armington model, the welfare consequences of changes in trade costs from τ
to τ 0 can be computed simply as:
1
b− ε
cj = λ
C
jj
b =
where, for any variable X, X
X0
X
(3.6)
denotes a proportional change in any variable between an initial
equilibria and a counterfactual one.
As the authors show, the only requirement to compute the changes in income are the knowledge of the
initial expenditure shares λij , the initial income levels Yi , and the trade elasticity ε. Those changes in
5
6
Here χij ≡ (Qi /ψij )ε
These results abstract from the dynamics involved in the change from one steady state to the other.
4
income levels can be used to compute the changes in the shares of expenditures for each possible trade
partner, λ̂ij and the changes in real consumption Ĉj .
As a preview of the counterfactual excercises that are performed, the simplest one that can be done
is comparing the current situation with autarky. A hypothetical autarky is whenever trade costs for
each pair of countries i 6= j approach infinity, τij → +∞ and there is no international trade between
countries. In that case, the algebra is simpler and the gains from trade, the absolute value of the
percentage change in real income associated with moving to autarky, is:
1
ε
.
Gj = 1 − λjj
3.2
(3.7)
Generalization of the model
CRC discuss how a single gravity equation can tie together many different types of models that are related
to different assumptions regarding the market structure: perfect competition, Bertrand competition,
and monopolistic competition with either homogeneous firms or firm-level heterogeneity. Those models
include, for example, Eaton & Kortum (2002), Bernard, Eaton, Jenson & Kortum (2000), Krugman
(1980), Chaney (2008), Arkolakis (2010), Arkolakis, Demidova, Klenow & Rodrı́guez-Clare (2008) and
Eaton, Kortum & Kramarz (2011).
In all of those models the assumption regarding preferences is a Constant Elasticity of Substitution
(CES) function
Z
σ−1/σ
Cj =
cj (ω)
σ/(σ−1)
dω
.
(3.8)
ω∈Ω
The models also assume balanced trade.7 In equilibrium a type of good is only imported from one
possible origin, so the consumption of goods produced in i in a destination j can be summarized as
Z
Cij =
!σ/(σ−1)
σ−1/σ
cj (ω)
dω
(3.9)
ω∈Ωij
where Ωij ∈ Ω is the set of goods that country j buys from country i, and the preferences are defined
over the continuum of goods Ω. The corresponding price index takes the same form as Equation (3.2),
and ψij = 1 is usually assumed for symmetry purposes. The price8 in country j of the goods imported
7
8
CRC show that allowing for trade deficits or surpluses increases the potential gains from trade.
This is an index of all the prices p(ω) of the goods ω ∈ Ωij exported from i to j.
5
from country i takes the form
!1−σ
Z
pj (ω)1−σ dω
Pij =
.
(3.10)
ω∈Ωij
In these models the set Ωij is an endogenous variable. Depending on the assumptions of each model
it is possible that some firms want to quit exporting to some destinations or producing at all. The
changes in the price index of the goods that are traded between a pair of countries reflect three different
elements: change at the intensive margin (change in the price of the goods) or change at the extensive
margin, either by the selection of a different set of firms that export from i to j or because of a different
set of firms producing at the origin i (entry).

Ej
× x
Pij =
cij
|{z}
Intensive margin
|
!
δ
1−σ
τij cpi
η
τij cpi 
×
Pj
{z
}
Extensive margin: selection
δ
Ri 1−σ
ce
| i {z }
×ξij
(3.11)
Extensive margin: entry
In this environment cpi , cei , cxij are variables that relate to variables costs of production, fixed entry
P
costs and fixed exporting costs, respectively. Ej ≡ ni=1 Xij is the total expenditure of country j,
P
while Ri ≡ nj=1 Xij is the total sales or revenues for producers from i. Also, ξij > 0 is a function of
structural parameters.
The most important parameters for this generalization are δ and η. The first one is a dummy variable
that takes a value of one with monopolistic competition with free entry. It takes a value of zero with
perfect or Bertrand competition. The parameter η ≥ 0 is related to the extent of heterogeneity across
varieties. It is related to the distribution from where the productivities of firms are assumed to take
their values. For example, in a monopolistic competition setup with fixed exporting costs like Krugman
(1980), it takes a value of η = 0, where the active firms always export, but it is η > 0 for other models
like Eaton & Kortum (2002), Chaney (2008), Melitz (2003) and others where there are firms that may
decide to stop exporting.
The other parameter, δ, is active in the model whenever there is monopolistic competition and free
entry. In those cases, a higher profitability of entry (low entry costs) in the domestic market causes
an increase in exporting varieties to all countries, which lowers the price of those varieties everywhere.
Therefore, parameters δ and η can be turned on and off, which allows for comparison of the results tied
to different assumptions of modeling.
6
3.3
One sector
In a one sector model where you hold the assumptions that the variable costs of production, fixed entry
e
costs, and fixed exporting costs are all used in the same proportions, cpi = cm
ii = ci = Yi , and trade in
goods is balanced, Ri = Yi , then CRC show that the price equation is
η
τij Yi 
ξij .
Pj
(3.12)
(Yi τij )−ε (cxij )−δη χij
P
Ej .
Xij = n
−ε x −δη χ
lj
l=1 (Yl τlj ) (clj )
(3.13)

Ej
Pij = τij Yi  x
cij
!
δ
1−σ
and the gravity equation is
In this case, ε = (1 + η)(σ − 1), so the interpretation of the trade elasticity is not the same as in the
1−σ
Armington model. Also, χij ≡ ξij
. In this case, the increase of trade costs affects both the price
of existing varieties (intensive margin) and the set of those varieties sold from country i to country j
(extensive margin). Even though there are differences in the model, the main contribution of Arkolakis
et al. (2012) was to show that the trade elasticity ε and the share of expenditure on domestic goods
λii remain the sufficient statistics for welfare analysis. In this case, the potential gains from trade are
the same as in the Armington model. This means that the separation of the intensive and extensive
margins does not change the magnitude of the gains from trade.
3.4
Multiple sectors
Multiple sectors s = 1, . . . , S can be incorporated into this setup by assuming that the preferences are
two-tiered. The upper level of the preferences is Cobb-Douglas in the form
β
Cj = ΠSs=1 Cj,sj,s ,
with βj,s ≥ 0 exogenous parameters and
PS
s=1 βj,s
(3.14)
= 1. The second tier is CES preferences that take
the following form:
Z
Cj,s =
cj,s (ω)
σs −1/σs
σs /(σs −1)
dω
,
(3.15)
ω∈Ω
7
and σs > 1 is the elasticity of substitution between different varieties, and can be different across
sectors. In that case, the price equation has to hold for each sector, and if the assumption that factors
of production are used in the same way across all activities in all sectors,9 and that the trade in goods
is balanced, Ri = Yi , the price equation can be expressed in the following form

Pij,s = τij,s Yi  ej,s
Ej
cxij,s
!
δs
1−σs
ηs
δs
τij,s Yi 
1−σs
ri,s
ξij,s ,
Pj,s
(3.16)
where ej,s ≡ Ej,s /Ej is the share of total expenditure in country j allocated to the sector s, and
ri,s ≡ Rj,s /Rj is the share of total revenues in country i generated from sector s. In this case there can
be monopolistically competitive sectors that cause scale effects through selection and entry. The latter
effect was not present in the one sector model because there ri,s = 1.
Here the gravity equation for each sector takes the form
Xij,s
3.5
δs
χij,s
(Yi τij,s )−εs (cxij,s )−δs ηs ri,s
P
ej,s Ej .
= n
−εs (cx )−δs ηs χ
lj,s
l=1 (Yl τlj,s )
lj,s
(3.17)
Tradable intermediate goods
Finally, tradable intermediate sectors can also be incorporated by assuming that in each sector the
production takes the form
Z
Ij,s =
ij,s (ω)
σs −1
σs
dω
σs
σs −1
(3.18)
ω∈Ω
Here, total expenditure equals total producer revenues, Ei = Ri , but the costs of production are allowed
to vary across sectors:
1−αi,s
cpi,s = Yi
α
ΠSk=1 Pi,ki,ks ,
where αi,ks are exogenous technology parameters that satisfy the condition that αi,s ≡
9
(3.19)
P
αi,ks . In this
e
This means cpi,s = cm
ii,s = ci,s = Yi .
8
case the price and gravity equations take the form

Pij,s
ej,s Yj
= τij,s ci,s 
vj cxij,s
!
δs
1−σs
ηs
τij,s ci,s 
Pj,s
and
Xij,s =
(τij,s ci,s )−εs (cxij,s )−δs ηs
ri,s Yi
vj ci,s
ri,s Yi
vj ci,s
δs
1−σs
ξij,s ,
(3.20)
δs
χij,s
ej,s Ej
Pn
rl,s Yl δs
x
−ε
−δ
η
s
s
s
χlj,s
(clj,s )
l=1 (cl,s τlj,s )
vl cl,s
(3.21)
where ci,s = cpi,s , vi ≡ Yi /Ri is the ratio of total income to total revenues in country i.
4
Counterfactual exercises
4.1
Gains from trade
All the versions of the model presented in Section 3 can be used to compute gains from trade using the
CRWIOD in a similar way as in the Armington model. For the Armington model we had a measure that
quantifies the changes in real consumption with respect to a change in the trade variables that in turn
affect the share of expenditure in domestic goods:
1
b− ε .
cj = λ
C
jj
(4.1)
The simplest counterfactual exercise that can be performed is to compute the changes in real income
with respect to autarky. This measurement gives an insight of how much a country gains from engaging
in international trade. In this case, the measure Gj quantifies the absolute value of the percentage
change in real income that would be associated with moving to autarky. For the cases with only one
sector,
1
ε
Gj = 1 − λjj
.
Given that λjj ≡ Xjj /Ej = 1 −
P
i6=j Xij /
(4.2)
X
, the numerator in the expression is simply total
ij
i=1
Pn
imports by country j, while the denominator is total expenditure by country j. The results presented in
this paper provide valuable information for two reasons. First, it updates the results from CRC from 2008
to 2011. Alessandria, Kaboski & Midrigan (2010), among others, discuss the magnitude of the trade
collapse that ocurred around the year 2009, which is resulting in lower gains from trade for all countries
9
and all possible specifications. The average gains from trade computed by CRC were 4.4%, and these
are 3.8% with the CRWIOD for 2011. Table 1 presents the results for the one sector models in column
2. These numbers use the number that CRC10 use for the elasticity in their baseline scenario ε = 5.
Eaton & Kortum (2002) find estimates from ε = 3.60 to ε = 12.86, and their preferred estimate is
ε = 8, 28. The aggregate elasticity can result in different magnitudes for the gains from trade, specially
when considering the sectoral elasticity as in Caliendo & Parro (2012). CRC discuss the sensitivity issues
regarding this parameter.
The second advantage of using this new database is that we can obtain measures for Costa Rica, which
was impossible before when the country was included in the Rest of the World data. Since Costa Rica
is very open to international trade, the estimated gains are relatively large. These gains are above the
average of the rest of the world and larger also than the rest of the world (the countries not included
in the database). The estimated gains of 5,0% are larger than those of other Latin American countries
such as Mexico (3.2%), Brazil (1.3%) and almost all other developed countries. The gains are, however,
smaller than other countries such as Ireland (8.3%), Belgium (6.6%) and Czech Republic (5.5%).
It is possible to obtain expressions equivalent to the latter for each version of the model presented in
the previous section. When multiple sectors are considered,
Gj = 1 −
ΠSs=1
λjj,s
ej,s
rj,s
δs !βj,s /εs
.
(4.3)
These gains from trade are on average larger than the ones computed for one sector models, such as the
ones presented in Arkolakis et al. (2012). Here, the scale effects play a role but the selection effects do
not. To estimate the gains from trade, the data from WIOD is used to compute the measures λjj,s , ej,s ,
βj,s and rj,s . Sector level trade elasticities εs are those from Caliendo & Parro (2012) for the sectors
included in manufacturing and agriculture, while for the sectors included in services, the assumption of
ε = 5 is kept. As I mentioned before, for sectoral computation using the CRWIOD we need to use 16
sectors instead of the 31 used in CRC because of how well disaggregated is the data included for Costa
Rica.
Table 1 shows in columns 3 and 4 the results for the different assumptions of competition schemes.
Costa Rica is one of the cases in which the gains in perfect competition are much larger than the ones
under the assumption of monopolistic competition, which occurs also in countries like Ireland. On the
opposite side, countries like Mexico and Brazil have larger gains under the monopolistic competition
assumption. This result depends on the comparative advantage of each country in the sectors that have
10
Further discussion can be found in Anderson & van Wincoop (2004).
10
strong scale effects. If this is the case, a country that opens to trade can specialize in sectors with large
scale effects. The opposite is true for countries like Costa Rica, in which specialization occurs in sectors
that do not have large scale effects.
When multiple sectors are included in the model the gains from trade increase dramatically. For the
Costa Rican case, they do so five times with respect to the benchmark case. The increase is significant
in part due to the Cobb-Douglas preferences assumed for the sectors. However, the increase is greater
in countries for which the market structure means that closing the possibility to trade may increase the
prices for some goods significantly, which is particularly sensitive with the Cobb-Douglas assumption.
Finally, when intermediate sectors are included in the model, the gains from trade can be computed
using the formula:
Gj = 1 − ΠSs,k=1
λjj,k
ej,k
bj,k
ηs
rj,k
bj,k
−δk !βj,s α̃j,sk /εk
,
(4.4)
P
where bj,k ≡ vj ( Sl=1 βj,l αj,kl ) and α̃j,sk is the elasticity of the price index in sector s with respect to
changes in the price index in sector k. The same variables λjj,s , ej,s , βj,s and rj,s are computed using
P
CRWIOD, which allows for computing the shares of intermediate purchases αj,ks = i Xij,js /Rj,s 11 .
Columns 5 and 6 show the potential gains from trade allowing for intermediates using two possible competition assumptions, perfect competition and monopolistic competition with firm level heterogeneity.
The gains are rather large compared to the previous models. This is mainly due to the fact that allowing
for trade in intermediates amplifies the gains of engaging in international trade. The welfare effects
occur in more than one round given that the price of intermediates used in production decreases in this
scenario.
Costa Rica is, again, one of the countries in which the gains from trade are smaller with the monopolistic
competition assumption. This is rather unusual when compared with the rest of the countries, given that
on average the gains increase from 23.2% in perfect competition to 33.4% in monopolistic competition
with the Melitz assumption. Again, this is the case for countries with comparative disadvantage in the
sectors with strong scale effects which are amplified in this setup.
11
However, for monopolistic competition, an additional assumption by Balistreri, Hillberry & Rutherford (2011) explained
in CRC must be made to obtain these measures. Otherwise, computation would be impossible.
11
Table 1: Gains from trade as percentages of income
Country
One Sector
Perf. Comp.
1
2
3
Costa Rica
AUS
AUT
BEL
BRA
CAN
CHN
CZE
DEU
DNK
ESP
FIN
FRA
GBR
GRC
HUN
IDN
IND
IRL
ITA
JPN
KOR
MEX
NLD
POL
PRT
ROM
RUS
SVK
SVN
SWE
TUR
TWN
USA
Rest of World
Average
5.0%
2.0%
4.8%
6.6%
1.3%
3.4%
1.5%
5.5%
3.8%
4.8%
2.5%
3.7%
2.6%
3.1%
3.6%
7.2%
2.2%
2.1%
8.3%
2.5%
1.3%
3.7%
3.2%
5.6%
4.0%
3.7%
3.8%
2.1%
6.5%
5.8%
4.6%
2.5%
5.5%
1.5%
3.7%
3.8%
23.6%
6.3%
23.1%
34.0%
3.2%
12.9%
2.4%
14.6%
11.8%
35.8%
6.9%
12.0%
8.1%
11.6%
16.4%
19.3%
4.2%
3.4%
20.3%
7.2%
1.3%
3.8%
10.5%
25.5%
15.1%
17.6%
11.0%
9.7%
18.1%
31.3%
10.9%
10.6%
7.5%
3.3%
11.6%
13.3%
Models
Multiple sectors
Mon. Comp.
Intermediates
Perf. Comp. Mon. Comp. (Melitz)
4
5
6
11.3%
2.6%
23.3%
33.2%
3.0%
11.7%
2.4%
18.9%
15.5%
30.7%
7.8%
11.6%
9.3%
10.6%
4.5%
21.3%
3.2%
3.6%
13.5%
7.6%
2.7%
8.1%
11.9%
23.6%
17.1%
13.7%
10.9%
0.7%
19.1%
32.8%
12.0%
10.6%
8.2%
3.1%
5.3%
12.1%
36.1%
11.7%
38.9%
55.3%
5.4%
22.4%
6.5%
31.9%
20.4%
52.4%
13.6%
20.8%
14.6%
20.2%
24.0%
38.4%
8.9%
6.7%
33.5%
13.2%
2.7%
10.2%
17.5%
41.5%
28.4%
29.0%
19.3%
16.9%
41.2%
50.8%
20.2%
18.5%
14.7%
5.8%
21.5%
23.2%
22.1%
3.6%
49.4%
67.8%
8.8%
32.1%
66.6%
74.3%
42.0%
51.1%
23.1%
27.2%
27.6%
21.5%
4.2%
67.8%
11.9%
10.5%
26.5%
18.7%
21.9%
74.3%
27.3%
46.3%
46.4%
26.5%
17.9%
-3.5%
79.8%
66.7%
33.5%
24.1%
28.2%
9.0%
15.0%
33.4%
Source: Computations using the Matlab programs used in Costinot & Rodrı́guez-Clare (2013). Data from CRWIOD for
2011. 16 sectors are used for computations with more than one sector. Trade elasticities from Caliendo & Parro (2012),
detailed in Appendix A.
12
4.2
Effects of increases in tariffs
It is possible to perform additional counterfactual exercises in the setup described. CRC show that
welfare changes given changes in tariffs in an Armington model can be computed using the following
formula:
1 − πj
1 − πj0
Ĉj =
where πj =
tij
i=1 1+tij λij
Pn
and πj0 =
!
−1/ε
λ̂jj
t0ij
i=1 1+t0ij λij λ̂ij
Pn
(4.5)
given that tij > 0 is the ad-valorem tariff
imposed by country j on goods imported from i. Graph 1 shows the potential gains (or losses) from a
country from unilaterally increasing the tariffs it charges for products entering the country. The potential
unilateral gains of a country like Costa Rica of increasing tariffs (assuming that other countries do not
retaliate) are larger than the average country, but not as large as those of a country that is even more
open to trade as Ireland.
Figure 1: Gains from a unilateral increase in the tariff
1.015
Costa Rica
USA
France
Portugal
Ireland
Welfare changes relative to 0% tariff regime
1.01
1.005
1
0.995
0.99
0.985
0.98
0.975
0.97
0
10
20
30
40
50
Tariff, %
60
70
80
90
100
Source: Computations using the Matlab programs used in Costinot & Rodrı́guez-Clare (2013). Data from CRWIOD 2011.
13
This is, however, an exercise that does not take into account the possibility of all countries reacting in
a similar manner and simultaneously increasing tariffs. On this issue, Table 3 takes one of the models
discussed in Section 3, and shows the effects of a simultaneous increase to a 40% tariff from the current
situation.12 In those cases, the magnitude of the potential losses is similar (in the opposite direction)
with the gains from trade when compared to autarky. This exercise must serve as a warning to evaluate
potential policy measures by one country. There are potential gains when acting on its own, but the
results from Table 3 show that an uncoordinated increase could cause a welfare loss for every single
country.
Table 2: Losses from a simultaneous 40% tariff increase.
Multiple sectors, intermediate goods, heterogeneous firms (Melitz)
Country
Costa Rica
AUS
AUT
BEL
BRA
CAN
CHN
CZE
DEU
DNK
ESP
FIN
Gj
-5.01%
-2.71%
-4.28%
-6.64%
-0.90%
-3.57%
-1.23%
-4.82%
-2.34%
-4.54%
-1.68%
-3.12%
Country
FRA
GBR
GRC
HUN
IDN
IND
IRL
ITA
JPN
KOR
MEX
NLD
Gj
-1.59%
-3.16%
-3.65%
-6.50%
-1.83%
-1.69%
-8.61%
-1.35%
-0.25%
-1.35%
-2.22%
-4.62%
Country
POL
PRT
ROM
RUS
SVK
SVN
SWE
TUR
TWN
USA
RoW
Average
Gj
-3.05%
-3.54%
-3.58%
-4.26%
-5.74%
-4.79%
-3.97%
-1.78%
-4.13%
-0.91%
-4.27%
-3.4%
Source: Computations using the Matlab programs used in Costinot & Rodrı́guez-Clare (2013). Data from CRWIOD for
2011. 16 sectors are used. Trade elastitcities from Caliendo & Parro (2012).
12
This is the multiple sectors, intermediate goods, heterogeneous firms (Melitz) version. The average loss from the one
sector model is 0,4%.
14
5
Conclusions
Costa Rica is a small open economy that benefits significantly from being open to trade. The magnitude of these gains and alternative trade policies can have a significant impact in the discussion of
alternative paths that a country can take deviating from the current situation. Given that the public
policy discussions surrounding these issues sometimes lack technical background, the results provided in
this paper are a reasonable starting point.
This paper computes the gains from trade using the methodology proposed by CRC and the new
CRWIOD database. The exercise using newer data for the countries that are part of WIOD allows for
updated results. The newer results imply lower gains, due in part to the collapse of international trade
in the aftermath of the international crisis of 2008-2009. Additionally, for the models with multiple
sectors, the gains could be lower when compared to CRC due to the fact that the computations could
only be done with fewer sectors.
The CRWIOD database also allows to compute the values for Costa Rica, which is not part of the
international version of the database. The results show that Costa Rica is one of the countries that gain
most from trade, and could have significant potential losses when compared to autarky or a simultaneous
increase of tariffs around the world. The gains are larger than the average of the world and are closer
to those of countries like Ireland than to those of large countries like the United States.
The results also show, as it was expected, that gains from trade are larger whenever the model includes
more sectors. However, for the Costa Rican case it is particularly interesting that those gains are
smaller in a monopolistic competition setup when compared to assumed perfectly competitive markets.
This result hints that the Costa Rican economy has less advantage in the sectors with the highest
scale effects. However, this characteristic is not unique to low and middle income economies, as other
advanced economies show this characteristic too.
Therefore, the results imply that Costa Rica is in a significantly better situation in terms of real income
thanks to its involvement in international trade. However, the advantages of the country may be hindered
by where they are located. Further research can be done to analyze how strategic public policies can be
pursued to increase the potential gain in sectors with greater scale effects.
15
References
Alessandria, G., Kaboski, J. P., & Midrigan, V. (2010). The Great Trade Collapse of 2008-09: An
Inventory Adjustment? NBER Working Papers 16059, National Bureau of Economic Research, Inc.
Anderson, J. E. & van Wincoop, E. (2004). Trade Costs. Journal of Economic Literature, 42 (3),
691–751.
Arkolakis, C. (2010). Market Penetration Costs and the New Consumers Margin in International Trade.
Journal of Political Economy, 118 (6), 1151 – 1199.
Arkolakis, C., Costinot, A., & Rodriguez-Clare, A. (2012). New Trade Models, Same Old Gains?
American Economic Review, 102 (1), 94–130.
Arkolakis, C., Demidova, S., Klenow, P. J., & Rodrı́guez-Clare, A. (2008). Endogenous Variety and the
Gains from Trade. NBER Working Papers 13933, National Bureau of Economic Research, Inc.
Balistreri, E. J., Hillberry, R. H., & Rutherford, T. F. (2011). Structural estimation and solution of
international trade models with heterogeneous firms. Journal of International Economics, 83 (2),
95–108.
Bernard, A. B., Eaton, J., Jenson, J. B., & Kortum, S. (2000). Plants and Productivity in International
Trade. NBER Working Papers 7688, National Bureau of Economic Research, Inc.
Bullón, D., Mena, T., Meng, B., Sanchez, N., Vargas, H., & Inomata, S. (2015). Using the inputoutput approach to measure participation in GVCs: The case of Costa Rica. IDE Discussion paper
529, Institute of Developing Economies.
Caliendo, L. & Parro, F. (2012). Estimates of the Trade and Welfare Effects of NAFTA. NBER Working
Papers 18508, National Bureau of Economic Research, Inc.
Chaney, T. (2008). Distorted Gravity: The Intensive and Extensive Margins of International Trade.
American Economic Review, 98 (4), 1707–21.
Costinot, A. & Rodrı́guez-Clare, A. (2013). Trade Theory with Numbers: Quantifying the Consequences
of Globalization. NBER Working Papers 18896, National Bureau of Economic Research, Inc.
Dietzenbacher, E., Los, B., Stehrer, R., Timmer, M., & de Vries, G. (2013). The Construction Of World
Input-Output Tables In The Wiod Project. Economic Systems Research, 25 (1), 71–98.
Eaton, J. & Kortum, S. (2002). Technology, Geography, and Trade. Econometrica, 70 (5), 1741–1779.
16
Eaton, J., Kortum, S., & Kramarz, F. (2011). An Anatomy of International Trade: Evidence From
French Firms. Econometrica, 79 (5), 1453–1498.
Krugman, P. (1980). Scale Economies, Product Differentiation, and the Pattern of Trade. American
Economic Review, 70 (5), 950–59.
Leiva, C. & Vargas, J. R. (2014). Estimación del Cambio Estructural por medio de los Modelos de
Insumo - Producto para Costa Rica de 1969 y 2011. Technical report.
Melitz, M. J. (2003). The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry
Productivity. Econometrica, 71 (6), 1695–1725.
Samuelson, P. A. (1939). The gains from international trade. Canadian Journal of Economics, 5 (2),
195–205.
17
Appendices
A
Description of the 16 sectors and trade elasticity used in estimation.
Table 3: Description of the 16 sectors used in estimation, and trade elasticity used.
Sector’s description
Agriculture, Hunting, Forestry and Fishing
Mining and Quarrying
Food, Beverages and Tobacco
Textiles and Textile Products, Leather, Leather and Footwear
Wood and Products of Wood and Cork
Pulp, Paper, Paper, Printing and Publishing
Coke, Refined Petroleum and Nuclear Fuel
Chemicals and Chemical Products
Rubber and Plastics
Other Non-Metallic Mineral
Basic Metals and Fabricated Metal
Machinery, Nec
Electrical and Optical Equipment
Transport Equipment
Manufacturing, Nec; Recycling
Electricity, Gas and Water Supply
Construction
Sale, Maintenance and Repair of
Motor Vehicles and Motorcycles; Retail Sale of Fuel
Wholesale Trade and Commission Trade, Except of
Motor Vehicles and Motorcycles
Retail Trade, Except of Motor Vehicles and
Motorcycles; Repair of Household Goods
Hotels and Restaurants
Inland Transport
Water Transport
Air Transport
Other Supporting and Auxiliary
Transport Activities; Activities of Travel Agencies
Post and Telecommunications
Financial Intermediation
Real Estate Activities
Renting of machinery and equipment and Other Business Activities
Education
Health and Social Work
Public Admin and Defence; Compulsory Social Security
Other Community, Social and Personal Services
Private Households with Employed Persons
16-sector
aggregation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Caliendo-Parro
trade elasticity
8,11
15,72
2,55
5,56
10,83
9,07
51,08
4,75
1,66
2,76
7,99
1,52
10,6
0,37
5
16
5
Source: Costinot & Rodrı́guez-Clare (2013) and Caliendo & Parro (2012).
18